Materials and Manufacturing Processes

ISSN: 1042-6914 (Print) 1532-2475 (Online) Journal homepage: https://www.tandfonline.com/loi/lmmp20

Experimental investigation on temperature-

affected magnetic abrasive finishing of aluminum
6060

Rajneesh Kumar Singh, Swati Gangwar & D.K. Singh

To cite this article: Rajneesh Kumar Singh, Swati Gangwar & D.K. Singh (2019)
Experimental investigation on temperature-affected magnetic abrasive finishing of
aluminum 6060, Materials and Manufacturing Processes, 34:11, 1274-1285, DOI:
10.1080/10426914.2019.1628263

To link to this article: https://doi.org/10.1080/10426914.2019.1628263

Published online: 25 Jun 2019.

Submit your article to this journal

Article views: 337

View related articles

View Crossmark data

Citing articles: 16 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=lmmp20
MATERIALS AND MANUFACTURING PROCESSES
2019, VOL. 34, NO. 11, 1274–1285
https://doi.org/10.1080/10426914.2019.1628263

Experimental investigation on temperature-affected magnetic abrasive finishing of

aluminum 6060
Rajneesh Kumar Singh , Swati Gangwar, and D.K. Singh
Mechanical Engineering Department, Madan Mohan Malaviya University of Technology, Gorakhpur, India

ABSTRACT ARTICLE HISTORY

Magnetic abrasive finishing (MAF) has a sound capability in boosting surface quality of advance material Received 6 February 2019
such as ceramics, composites, and hard alloys. The surface finish and quality have a greater dependency Accepted 30 May 2019
on the rise of temperature of the target surface during the microfinishing via MAF process. In present KEYWORDS
study, experiments were done to measure the “rise in temperature” of aluminum 6060 microfinished Magnetic; abrasive; finishing;
surface. Taguchi “L9 Orthogonal Array” (L9 OA) was employed to design the experiments with voltage, aluminum 6060; Taguchi; L9
abrasive weight, working gap, and rotational speed opted process parameters. Buckingham π-theorem orthogonal; array;
was used to formulate a semiempirical dimensional model to forecast the “rise in temperature.” The temperature; Buckingham-π
dimensionless constant and exponents were calculated by nonlinear approximation of L9 orthogonal
array experimental results. The dimensionless constant (K) of predictive model was 6.3066 × 10−6 and
exponents (ϕ,φ,γ, and δ) were calculated 1.7509, −0.6651, 0.3934, and −0.0830, respectively. Finally, to
explore the flexibility of developed Buckingham-π dimensional model to other materials, this model was
applied on mild steel workpiece to predict and analyze the raise in temperature microfinished using
MAF process. Predictive semiempirical temperature model was validated with confirmatory test results.
The predictive results had good agreement with 7.31% average error in respect to test results.

Introduction finish of the internal MAF process and through scanning

electron microscopy (SEM), the microscopic surface texture
MAF process is an advanced finishing operation that was intro-
of workpiece was examined.[16] Design of flat faced electro-
duced to finish unconventional engineering materials such as
magnet was done by Singh et al. for finishing flat surfaces.
ceramic, composite, and super-alloys.[1,2] This process had
Their research contributed in finding the most substantial
grown into a sophisticated finishing operation, producing impec-
input parameters that affects the surface finish were “voltage
cable surface texture without altering surface integrity.[3–6] In,
and machining gap,” trailed by “grain mesh and rotational
MAF process magnetic field is used to control the magnetic
speed.”[17] Singh et al. clarified role of Flexible-MAB on sur-
abrasive particles (MAPs) in the working gap to remove the
face texture of workpiece created by abrasive particles after
material from the target surface. MAPs were created by mixing
the finishing of processed surface by microscopic study.[18]
abrasive particles with larger ferrous particles in proper propor-
Recently, Ahmad et al. optimized the significant process para-
tions. MAPs create a magnetic abrasive brush (MAB) under the
meters, that contributes in improving the surface roughness of
effect of magnetic field and acted as multipoint cutting head
stainless steel 202 finished through MAF process. They
capable in dislodging the material in-form of micro-to-nano
emphasized to reveal basic characteristics of sintered MAPs.
chips.[7–10] A substantial amount of heat was produced due to
In their MAPs, Al2O3 abrasive particles was stacked on the
plastic deformation and rubbing between brush-finishing surface
ferrous particles served as base metal, eventually improving its
during finishing.[11–14] Excessive temperature lead to surface con-
quality.[19]
tamination by fusing abrasive particles in the finishing zone and
Hou and Komanduri used “Jaeger” standard moving heat
shows the thermal strain hardening and corrosion tendency might
theorem to develop an analytical thermal formulation to ana-
increase due to alteration in chemical composition of workpiece
lyze effect of temperature on the finished surface. Their for-
surface due to presence of excessive temperature.
mulation effectively calculated flash temperature and flash
Shinmura et al. constructed dedicated MAF set-up for
time that generated during finishing. They developed
finishing outer surface of cylindrical parts by retrofitting
a general solution for both “transient and steady state” using
conventional lathe machine with two poled electromagnets.
moving ring and disc heat source equations.[20] Kumar and
Flexible-MAB was formed when MAPs aligned with magnetic
Yadav used Finite element method to investigate the basic
forces in the machining gap present between cylindrical part
characteristics of temperature on the microfinished surface.
and electromagnet, which enhanced the surface finish of outer
They concluded that raise in temperature was an effect of
surface cylindrical workpiece.[15] Yamaguchi and Shinmura
MAPs velocity and magnetic intensity. To reduce the surface
investigated the impact of process parameters on surface

CONTACT Swati Gangwar [email protected] Mechanical Engineering Department, Madan Mohan Malaviya University of Technology, Gorakhpur 273010, India
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/lmmp.
© 2019 Taylor & Francis
 MATERIALS AND MANUFACTURING PROCESSES 1275

temperature, tool speed, and magnetic flux density (MFD) approximately 6%. Further, their key findings were tempera-
must be wisely determined.[21] To measure surface tempera- ture depends on abrasive weight, voltage, and pulse on time.
ture of workpiece during finishing process, many studies were Maximum temperature was raised to 46°C which showed that
done on different temperature measuring methods. Xu and it was a low-temperature finishing process. Ferrous alloys are
Malkin compared embedded thermocouple and two color good thermal conductor and it was also seen that temperature
infrared detector measuring techniques to measure subsurface rise was not high as it was seen in case of ceramics.[27]
temperature and foil thermocouple to measure surface In present study, experiments were executed on MAF setup
temperature.[22] Similarly, Kato and Fujji used physical to obtain the “rise in temperature” between flexible magnetic
vapor deposition (PVD) to measure temperature of plain abrasive brush and finishing surface aluminum 6060 surface.
carbon steel and stainless steel but they observed the creation Taguchi L9 OA was applied for systematic designing of experi-
of a thin film of vapor and placing on the surface of workpiece ments. Process parameters selected were abrasive weight, vol-
was a tedious job.[23] Ueda et al. developed noncontact fiber tage, working gap, and rotational speed of electromagnet.
coupled pyrometer to measure temperature during end Experimental results were used to develop a semiempirical
milling operation.[24] Hence, most suitable method to mea- model based on Buckingham-π theorem to predict the rise in
sure surface temperature was use of the thermocouple. They temperature. Validation of semiempirical model was done by
are part of contact type measuring methods and largely pre- comparing it with confirmatory tests performed by taking
ferred by researchers to measure the temperature in finishing different processing conditions from main L9 experiments.
process. Finally, analysis of semiempirical model was done to analyze
Further, for incorporating the effect of workpiece material the behavior of process parameters and properties of alumi-
properties on the response performance of machining and num 6060 in raising the surface temperature. This model had
finishing process, many investigations were done using agreeably revealed the effect of process parameters and work-
dimensional analysis method. Dimensional analysis, an piece material properties contributing in raising the surface
accomplished method in formulating and predicting the beha- temperature. Eventually, the developed Buckingham-π dimen-
vior of material properties. It is generally based on sional model was implemented on mild steel workpiece to
Buckingham’s π theorem which is used to derive predict the raise in surface temperature microfinished via
a semiempirical equation for investigation into the effects of MAF process.
material properties of workpiece. Tsai and Wang did
a dimensional analysis of surface roughness in EDM and
compared developed model with nonlinear optimization Materials and methods
method. They revealed that surface finish was depended on
the material properties and input parameters, that is, electric Magnetic abrasive finishing setup
polarity and peak current.[25] Patil and Brahmankar applied MAF setup was fabricated in-house, consisted of plane-faced
dimensional modelling to determine material removal rate of two pole electromagnetic tool mounted on pillar drilling
metal matrix composite machined by using Wire-EDM. In machine, that is, shown in Fig. 1. Pillar drilling machine had
their study, coefficient of thermal expansion played rotational speed between “100 and 300 rpm” on which elec-
a dominant role.[26] Mulik et al. did experimentation to tromagnetic tool was mounted. The electromagnetic tool con-
study the behavior of temperature on the finishing surface sists of 1000 turns of 18-gauge copper wire winding wrapped
of ferrous alloy via Ultrasonic-MAF process. Buckingham-π over a mild steel core (ϕ30 mm). The whole core-coil winding
theorem was applied to model temperature generation during was sealed into 5 mm mild steel casing for reducing magnetic
finishing of workpiece. Their predictive model had a decent flux leakage into surrounding space. A regulated DC power
accordance with experimental results with an error of bridge rectifier was designed and fabricated to provide

Figure 1. Magnetic abrasive finishing experimental setup.

1276 R. K. SINGH ET AL.

constant power supply to electromagnet. The operating range flow. MFD in the range of “0–1.02 T” was produced in the 1 mm
of DC power supply was between “0 and 24 V” voltage and “0 working gap with regulated DC supply of “3.0–6.0 A.”
and 10 A” current. DC supply was fed to the electromagnetic
tool with help of slip ring and carbon brush assembly. This
slip ring enables seamless direct current supply to the rotating Temperature measurement
electromagnetic tool. The regulated direct current energized
the electromagnetic tool and a strong magnetic field in the K-type thermocouple was used for measuring temperature of
working gap was produced. aluminum 6060 in present surface temperature experiments.
Chromel and Alumel were the main combination of K-type
thermocouple. Accuracy of this thermocouple was ±1.0°C and
had the temperature range between −50°C to +750°C. Shen
Magnetic flux density measurement
et al. developed a fixing methodology to measure surface
The strength of magnetic flux density (MFD) was measured by temperature using a thermocouple in the grinding
EMF-Portable gaussmeter with a measuring range “0.0 − 2.0 T.” process.[28] Figure 3 shows the surface temperature arrange-
Probe of digital gaussmeter was moved radially inward from ment used for present study, three ϕ 2 mm through hole at
edge of outer casing to center of electromagnet into working the specific location were drilled into workpieces. Then
gap to measure the MFD. Figure 2. display the variation of MFD K-type thermocouples with ϕ 1.6 mm probe were inserted
in working gap at the different position. MFD was highest in the in the holes. Positioning of holes were 15 mm apart from each
core of electromagnetic tool and lower at the outer casing of other decided by strength of MFD in the working gap. The
electromagnetic tool. Mild steel core had more volume then the surface temperature was recorded in the digital temperature
casing that enable more magnetic material for magnetic flux to indicators (T1, T2, T3).

1.2

3.0 A Current
1
Magnetic flux density(T)

4.5 A Current
0.8 6.0 A Current

0.6

0.4

0.2

0
0 5 10 15 20 25 30 35 40 45 50
Radial distance of magnet from centre (mm)

Figure 2. Magnetic flux distribution along radial distance.

Figure 3. Method of insertion of k-type thermocouple probe.

 MATERIALS AND MANUFACTURING PROCESSES 1277

Experimental procedure Table 2. L9 orthogonal array for surface temperature.

ΔRa “Rise in temperature” (°C)
The aluminum 6060 was opted as workpiece material with
Trial no. P1 P2 P3 P4 (µm) Initial Final Δ°C
Vicker’s hardness “31–34 HV” for present study. The dimension
1 1.0 20 10 100 0.15 23 26 3.0
of the workpiece was “100 × 100 × 10 mm3,” where 10 mm was 2 1.0 25 14 200 0.19 23 28 5.0
depth of workpiece. Primarily, all the workpieces were prefinished 3 1.0 30 18 300 0.23 23 32 9.0
4 1.5 20 14 300 0.17 23 27 4.0
to remove surface irregularities and defects. Then, surface rough- 5 1.5 25 18 100 0.21 23 30 7.0
ness measurement was done by Ziess-Handysurf surface tester 6 1.5 30 10 200 0.18 23 28 5.0
with resolution 0.01 µm at ±20 µm and surface roughness (Ra) 7 2.0 20 18 200 0.19 23 28 5.0
8 2.0 25 10 300 0.16 23 26 3.0
value of prefinished was found between “1.43 and 1.56 µm.” The 9 2.0 30 14 100 0.17 23 27 4.0
Ra was measured at five different position on each sample and
a mean Ra value was obtained. The measurement of Ra was done
at the same position before microfinishing and after microfinish- the selected input parameters and “rise in temperature” as output
ing of workpiece. The experiments started by mounting work- response.[29] Surface temperature experiments were conducted
piece (inserted K-type thermocouples) on the worktable of pillar using Taguchi L9 orthogonal array with different processing con-
drill machine. ditions shown in Table 2. The room temperature was 23 °C. After
Dial gauge was used to level the workpiece and workpiece completing L9 experiments, it was observed that maximum sur-
centering was done along with the center of electromagnetic face temperature of aluminum 6060 was recorded at 32°C. Hence,
tool. Slips gauges were used to measure specified working gap maximum “rise in temperature” of aluminum 6060 surface was
between finishing surface of workpiece and electromagnetic tool. found 9° C for trial no 3. After the completion of experiments,
Unbonded type-MAPs (UMAPs) mixture were made by mixing final surface roughness of microfinished surface was measured
ferrous particles (Fe, 300 mesh size) with silicon carbide (SiC, 400 and change in surface roughness (ΔRa) of aluminum 6060 are
mesh size). SiC was mixed in essential proportion, that is, 20 g, shown in Table 2. Figure 4 shows the SEM image of change in
25 g, and 30 g in iron powder to form three 100 g samples of surface roughness of aluminum 6060 for trial condition 3. Figure 4
UMAPs. Then, required amount of UMAPs were filled into (a) shows the grinding marks after grinding process. Figure 4(b)
working gap “1–2 mm.” These UMAPs formed the FMAB when displays the surface texture produced by the FMAB of MAF
the DC supply was fed to the electromagnet. FMAB removed the process. The SEM imagery revealed that the microfinishing of
material from the workpiece surface when rotational speed aluminum 6060 surface via MAF process was done by scratching/
“100–300 rpm” was introduced. Selected process parameters for microcutting. The change in surface roughness was found to be
performing the experiments on aluminum 6060 are shown in high with respect to higher values of voltage and abrasive weight.
Table 1. The range of these process parameters were determined Experimental results showed that MAF process generate a very
as per fabricated in-house MAF setup constrains. Then, pilot low temperature during microfinishing of aluminum 6060 which
experiments were done to examine the proper functioning of makes it very thermal stable process along with good surface
the fabricated MAF setup, and it observed that “rise in tempera- finish.
ture” became stable approximately around 20 min.
The Taguchi L9 orthogonal array experiment design was
employed for systematic surface temperature experiments using Dimensional analysis by Buckingham-π theorem
Dimensional analysis is a sound method to develop a predictive
semiempirical model that incorporates process variables along
Table 1. Selected process parameters range.
with significant physical properties of system. This method is
Sr no Process parameters Units Range
capable to model various machining and finishing process.
1 Working gap (P1) mm 1.0–2.0
2 Abrasive weight (P2) g 20–30 Langhaar was the pioneer who expand the scope of dimensional
3 Voltage (P3) V 10–18 analysis using in different engineering field such as strength of
4 Rotational speed (P4) rpm 100–300 material, theory of heat, electromagnetism, and differential

Figure 4. SEM image: (a) prefinished surface and (b) postfinished surface.
1278 R. K. SINGH ET AL.

equations.[30] Dimensional analysis is a method which reduces Table 4. Thermo-physical properties of aluminum 6060.
the complexity of any physical and experiment problem. By Characteristic Factors Symbols Dimensions Value Units
reducing complex problem from n-dimensional variable to Material Density of ρ ML−3 2700 kg/m3
q dimensionless variables and the reduction shows, n-q = 1,2,3 properties material
Thermal k MLT−3θ−1 209 W/m-K
depending on problem complexity. Generally, n-q equals num- conductivity
2 −2 −1
ber of physical dimensions which rules the problem. The major Specific Cρ LT θ 0.91 J/g°K
heat
advantages of dimensional analysis are, problem is reduced to capacity
−1
a lesser number of dimensionless variables which is lesser than Thermal α θ 23.4 K−1
the main physical variables and quantitative prediction of expansion
Electrical ρe ML3T−3Q−2 0.032 × 10−6 ohm-m
a mathematical relationship between physical variables, para- resistivity
meters, and constant of a physical event becomes easy.
Dimensional analysis is a powerful method which can
develop semiempirical predictive model by considering the physical variables are stated in form of dimensionless vari-
important physical properties of the workpiece material. ables (πs) and the problem will reduce from the original n to
Empirical methods such as regression analysis, Taguchi ortho- n-q, where n are variables of physical quantities and q are
gonal design, and response surface methodology are some com- fundamental dimensional parts. Buckingham’s π theorem was
mon methods used to develop a predictive model, but they only applied to formulate dimensionless products by using vari-
considered the machining process parameters. These methods ables present in above problem statement. A dimensional
totally neglect the consequence of physical properties of the matrix was then framed based on basic physical quantities
workpiece on response performance of the machining process. such as “mass (M), length (L), time (T), temperature (θ), and
To incorporate the advantages of dimensional analysis and to charge (Q)” are shown in Table 5, where all dependent vari-
develop a predictive model considering process parameters ables were specified as per their fundamental dimensions,
along with physical properties of workpiece material, this pre- where “z1, z2, z3, z4, z5, z6, z7, z8, z9, and z10” were the indices
sent study was performed. For development of the predictive of variables in Eq. (1).
model of the rise in temperature, major process parameters The Buckingham’s π theorem tells that, complete physical
considered were voltage, working gap between workpiece and equations must be dimensionally homogeneous and can
electromagnet tool, abrasive weight in UMAPs and rotational reduce the number of independent quantities in a problem
speed of the electromagnet. All important details of process by q and solution can obtained by the product of independent
parameters which were important in the development of the dimensionless parts πn-q, where n is variables present in
predictive model along with response measure, that is, tempera- dimensional equations and q is rank of the dimensional
ture rise, are listed in Table 3. matrix and final equation can be written as follows:
The main physical properties of the workpiece were den-
sity, thermal conductivity, thermal linear expansion, specific

f π1; π2; π3 : : : : : : πnq ¼ 0 (2)
heat capacity, and electrical resistivity. The details of physical
properties that were important in development of the predic- By applying Buckingham-π theorem, the dimension formula
tive model are listed in Table 4. These properties were con- for Eq. (1) was derived by placing the fundamental values of
sidered due to the “rise in temperature” depends on the different factors in Eq. (2) and derived relation is as follows:
thermo-physical properties of material.
The predictive “rise in temperature” model of MAF process  Z2  Z4  1 Z5  3 Z6
was developed by using following expression[27]: ½θZ1 ML2 T 3 Q1 ½LZ3 MLT 2 T ML
     
Z7 2 2 1 Z8 1 Z9  Z10
MLT 3 θ1 LT θ θ ML3 T 3 Q2 (3)
T ¼ f ðV; L; w; N; ρ; k; Cρ; α; ρe Þ (1)  
¼ M 0 L0 T 0 θ0 Q0
Where T is “rise in temperature” which is the response
measure of MAF process. V is voltage, L is working gap, w is Based on homogeneous linear equations for dimension pro-
SiC abrasive weight in MAPs, N is rotational speed are process ducts, above dimensional Eq. (3) was rearranged in form of
parameters of the MAF, and ρ is density of aluminum 6060, Table 5 and further dimensional formula was written inform
Cρ is specific heat capacity, k is thermal conductivity, α is of equations by using the coefficients of the dimensional
linear thermal expansion, and ρe is electrical resistivity are the products given in the Table 5.
physical properties of workpiece material.
Buckingham developed dimensional theorem using physi-
cal quantities. He stated that, when whole dimensional Table 5. Dimension of physical properties and parameters.
Factors
Table 3. Process parameters of MAF process. Z1 Z2 Z3 Z4 Z5 Z6 Z7 Z8 Z9 Z10
Characteristics Factors Symbols Dimensions Units Dimensions (T) (V) (L) (w) (N) ðρÞ (k) ðCρÞ (α) ðρe Þ
Response Temperature rise T θ °C M 0 1 0 1 0 1 1 0 0 1
Process parameters Voltage V ML2T−3Q−1 volt L 0 2 1 1 0 −3 1 2 0 3
Working gap L L mm T 0 −3 0 −2 −1 0 −3 −2 0 −3
Abrasive weight w MLT−2 g Θ 1 0 0 0 0 0 −1 −1 −1 0
Rotation speed N T−1 rpm Q 0 −1 0 0 0 0 0 0 0 −2
 MATERIALS AND MANUFACTURING PROCESSES 1279

9
Dimensional predictive model z6 ¼ z3 þ z4  z5 >
>
z7 ¼  12 z2  z3  2z4 þ z5 >
>
=
The number of dimensionless products (πs) in the dimen-
z8 ¼ 2 z3 þ 2z4  2z5
3
(10)
sional model could be obtain by two methods; either by >
>
z9 ¼ z1 þ 12 z2  12 z3 þ z5 >
>
finding rank of dimensional matrix formed by the last five ;
z10 ¼  2 z2
1
columns the dimension table of parameters and properties,
that is, Table 5[21] or by simply subtracting number of funda- Putting the value of z1 ¼ 1; z2 ¼ z3 ¼ z4 ¼ z5 ¼ 0 in
mental dimensions (q) from total physical variables (n) of the Eq. (10), gave first solution as:
homogeneous dimensional model.[22] In present case both z6 ¼ 0; z7 ¼ 0; z8 ¼ 0; z9 ¼ 1, and z10 ¼ 0.
methods were opted to verify whether both the methods Putting the value of z2 ¼ 1; z1 ¼ z3 ¼ z4 ¼ z5 ¼ 0 in
provide same solution or not. Eq. (10), gave second solution as:
Firstly, number of dimensionless products (πs) was z6 ¼ 0; z7 ¼  12 ; z8 ¼ 12 ; z9 ¼ 0, and z10 ¼  12 .
obtained by determining the rank of matrix obtained by Putting the value of z3 ¼ 1; z1 ¼ z2 ¼ z4 ¼ z5 ¼ 0 in
forming a square matrix of last five columns of Table 5. Eq. (10) gave third solution as:
Rank was determined by solving determinant of square matrix z6 ¼ 1; z7 ¼ 1; z8 ¼ 32 ; z9 ¼  12 , and z10 ¼ 0.
of fifth order and determinant was −4, that is, not equal to Putting the value of z4 ¼ 1; z1 ¼ z2 ¼ z3 ¼ z5 ¼ 0 in
zero. Hence rank of matrix was determined to be 5, as given Eq. (10), gave forth solution as:
below. Therefore, number of dimensionless products (πs) z6 ¼ 1; z7 ¼ 2; z8 ¼ 2; z9 ¼ 0, and z10 ¼ 0.
were taken as the difference between total variables (n = 10) Putting the value of z5 ¼ 1; z1 ¼ z2 ¼ z3 ¼ z4 ¼ 0 in
and derived rank of matrix (q = 5), that is, 10–5 = 5. Eq. (10), gave fifth solution as:
  z6 ¼ 1; z7 ¼ 1; z8 ¼ 2; z9 ¼ 1, and z10 ¼ 0.
 1 1 0 0 1 
 All power indexes values of dimensional parameters that
 3 1 2 0 3 
 was derived by solving Eq. (10) are listed in the Table 6. There
 0 3 2 0 3  ¼ 4; ) matrixrank : 5
  were total 10 fundamental variables (n = 10) and total 5
 0 1 1 1 0 
  fundamental dimensions (q = 5). Dimensionless solution
 0 0 0 0 2  matrix contained five rows (i.e., n-q = 5) in this case. Each
row in the solution matrix contained an exclusive dimension-
Then, the second method was applied to determine the num-
less product. For the present case dimensionless product are
ber of dimensionless products (πs) and was obtained by sub-
tracting number of fundamental dimensions (q) from physical
as follows: pffiffi 3=2
V α pffiffi
2
π1 ¼ Tα, π2 ¼ p ffiffiffiffiffi , π3 ¼ LρCρ
k α
, π4 ¼ wρCρ
k2 , π 5 ¼ ρCρ2
Nkα
variables (n) in the dimensional model. There were 10 physi- kρe
cal variables and 5 fundamental dimensions (M, L, T, θ, and The above relations between different dimensionless pro-
Q) in the present case. Hence, dimensional model was ducts was arranged as:
reduced to 10–5 = 5 dimensionless products (πs).

π1 ¼ f π2 ; π3 ; π4 ; π5
As, it was seen that both the methods provide the same
solution for determining the number of dimensionless pro- or
pffiffiffi !ϕ  φ  γ 
ducts (πs), that is, 5, Hence five dimensionless products (πs) 1 V α LρCρ3=2 wρCρ2 Nkα δ
are present in the whole set of equations of dimensional T ¼ K pffiffiffiffiffiffiffi  pffiffiffi  
α kρe k α k 2 ρCρ2
model.
(11)
f ðπ1 ; π2 ; π3 ; π4 ; π5 Þ ¼ 0 (4)
Where K is dimensional constant and ϕ,φ,γ and δ are
By equating power of fundamental units of both side of Eq. unknown exponents of the dimensionless products. The
(3), a set of linear Eqs. (5–9) were established and then most significant physical properties on which this semiempi-
magnitude of constants was obtained by solving Eqs. (5–9) rical model depended on thermal conductivity and linear
simultaneously. thermal expansion of workpiece material. The dimensional
temperature rise model depended on both process parameters
z2 þ z4 þ z6 þ z7 þ z10 ¼ 0 (5) and physical properties of workpiece. Value of dimensionless
constant K and values of exponents ϕ,φ,γ, and δ were founded
2z2 þ z3 þ z4  3z6 þ z7 þ 2z8 þ 3z10 ¼ 0 (6)
Table 6. Dimensional analysis results.
 3z2  2z4  z5  3z7  2z8  3z10 ¼ 0 (7) Factors
Z1 Z2 Z3 Z4 Z5 Z6 Z7 Z8 Z9 Z10
z1  z7  z8  z9 ¼ 0 (8) (T) (V) (L) (w) (N) ðρÞ (k) ðCρÞ (α) ðρe Þ
π1 1 0 0 0 0 0 0 0 1 0
 z2  2z10 ¼ 0 (9) π2 0 1 0 0 0 0 −1/2 0 1/2 −1/2
π3 0 0 1 0 0 1 −1 3/2 −1/2 0
π4 0 0 0 1 0 1 −2 2 0 0
Solving equations from 5 to 9 for value of z6 ; z7 ; z8 ; z9 , and π5 0 0 0 0 1 −1 1 −2 1 0
z10 , we got:
1280 R. K. SINGH ET AL.

by nonlinear approximation of experimental results which of aluminum 6060 is 209 W/m-K, that is, fairly high, hence it
was performed earlier and their results are given in Table 2. became easy to conduct heat rapidly from the microfinishing
Values of K,ϕ,φ,γ, and δ were 6.3066 * 10−6, 1.7509, −0.6651, interface surface between FMAB-workpiece. Therefore, sur-
0.3934, and −0.0830, respectively were obtained by writing face temperature during microfinishing does not rise to a high
a mathematical code in MATLAB software; version R2014a temperature, making microfinishing on aluminum 6060 sur-
using the above-developed dimensional equations. The final face via MAF process is low heat generation process.
dimensional solution product is given in Eq. (12): Further, aluminum 6060 has 23.4 K−1 linear thermal
pffiffiffi !1:7509 expansion coefficient which make it easy to perform low
1 6 V α thermal microfinishing operation without any linear thermal
T ¼ ð6:3066  10 Þ  pffiffiffiffiffiffiffi
α kρe expansion. After the thermal conductivity and linear thermal
 0:6651  0:3934 expansion, “rise in temperature” subsequently depends on the
LρCρ3=2 wρCρ2 aluminum 6060 specific heat capacity. The specific heat capa-
 pffiffiffi 
k α k2 city has the repeatability in three-dimensional products.
 0:0830
Nkα Specific heat capacity of aluminum 6060 is 0.91 J/g°K that
 (12)
ρCρ2 need higher temperature to raise the surface temperature of
workpiece. The maximum surface temperature of aluminum
Equation (12) of dimensionless product was used to predict 6060 was 32°C during microfinishing which is insignificant.
the “rise in temperature” during the surface microfinishing of So, it makes aluminum 6060 more thermal stable material to
aluminum 6060 workpiece via MAF process. “rise in tempera- be finished with MAF process.
ture” was predicted by placing all value of process parameters
and material properties in the Eq. (12).
Parametric analysis
Results and discussion
Parametric analysis using developed dimensional model
Dimensional modelling is sophisticated semiempirical model- revealed the significant behavior of process parameters of the
ling method that includes material properties of opted work- MAF process. Interactive behavior of process parameters of
piece along with process parameters of finishing process. MAF process affecting temperature rise on surface of alumi-
Developed dimension model for the predicting “rise in tem- num 6060 are shown in Fig. 5. The exponents of dimensional
perature” was based on L9 experimental results. The “rise in product affect the behavior of process parameters, a positive
temperature” on the finishing surface was significantly low as exponent gives the direct relationship between process para-
revealed by experimental results. The dimensional model pro- meters and output response and a negative exponent of dimen-
vided the reasons for such low “rise in temperature” on the sional product give inverse relationship. Further, significance
finishing surface of aluminum 6060 finished via MAF process. of process parameters depends on the value of exponents of
This model reveals the contribution of material properties in dimensional product, higher the value more the significance of
raising the surface temperature along the process parameters process parameters. The mutual effect of process parameters on
of MAF process. Effect of process parameters and significant the “rise in temperature” was analyzed.
material properties on which surface temperature rise Figure 5(a) shows the influence of voltage on the “rise in
depends are analyzed in detail. temperature” in interaction with abrasive weight. With increase
in voltage, “magnetic flux density” in working gap also increases.
FMAB becomes more rigid with rise in “magnetic flux density”
Effect of material properties
leading to rise in the magnetic normal force. Magnetic normal
The thermal conductivity, thermal linear expansion, specific force increases the penetration of magnetic abrasive particles
heat capacity of workpiece material plays vital role on the into workpiece surface. Therefore, the interaction between work-
“rise in temperature” on the microfinished surface. The sig- piece surface with abrasive particles increases resulting in
nificance of these material properties on which surface tem- increases in the “rise in temperature.” Figure 5(b) shows the
perature rise depends can be found out by analyzing impact of SiC abrasive weight in magnetic abrasive brush on the
dimensional products. Semiempirical dimensional model con- “rise in temperature” in interaction with rotational speed. With,
sists of five-dimensional products considering the five an increase in abrasive weight more abrasive particles are avail-
thermo-physical properties of aluminum 6060 workpiece. able to erode the workpiece surface. Therefore, more friction is
Repeatability of the thermo-physical properties in the dimen- generated due to more interaction between abrasive particles and
sional products helps in revealing their significance and con- surface, which increases the surface temperature. Hence leads to
tribution in raising surface temperature. “rise in temperature” on aluminum 6060 finishing surface.
Dimensional model revealed that “rise in temperature” Figure 5(c) shows the basic behavior of working gap on the rise
highly depend on the thermal conductivity and linear thermal in temperature in interaction with voltage. It is observed with rise
expansion of aluminum 6060. Equation (12) of dimensional in working gap, density of magnetic flux declines in working gap.
model depicts that significance of any material property Therefore, strength of magnetic abrasive brush decreases which
depends on the repeatability of material property in the reduces the force on the magnetic particles and a lesser “rise in
dimensional products, that is, four in case of thermal con- temperature” due to low friction produced due to lower interac-
ductivity and linear thermal expansion. Thermal conductivity tion between abrasive particles and workpiece surface. Figure 5(d)
 MATERIALS AND MANUFACTURING PROCESSES 1281

Voltage-Abrasive weight Abrasive weight-Rotational speed

12
4.5
4

Rise in temperature (oC)

10

Rise in temperature (oC)

3.5
8 3

6 2.5
2
4 100 rpm
SiC 20g 1.5
200 rpm
SiC 25g
2 1 300 rpm
SiC 30g
0.5
0 0
10 14 18 20 25 30
Voltage (V) Abrasive weight (g)
(a) (b)
Working gap-Voltage Rotational speed-Working gap
12 6

Rise in temperature (oC)

10 5
Rise in temperature (oC)

8 4

6 3

4 2
1.0 mm 100 rpm
1.5 mm 200 rpm
2 1
2.0 mm 300 rpm
0 0
10 14 18 1 1.5 2
Voltage (V) Working gap (mm)
(c) (d)

Figure 5. Characteristics plots of process parameters with rise in temperature: (a) influence of voltage and abrasive weight; at working gap = 1 mm and rotational
speed = 100 rpm, (b) influence of abrasive weight and rotational speed; at voltage = 14 V and working gap = 2 mm, (c) influence of working gap and voltage; at
abrasive weight = 30 gm and rotational speed = 100, (d) influence of rotational speed and working gap; at abrasive weight = 20 gm and voltage = 14.

shows the effect of rotational speed of the electromagnet on the and predictive results for all five trials. It was found that dimen-
“rise in temperature” in interaction with working gap. This hap- sional model was highly accurate and closest predicted was at trial
pen due velocity of abrasive particles become more which no. 3 with an error of 2.03%. Poorest prediction was at trial no. 5
increases the rate of toppling of abrasive particles in FMAB. with 13.82.%. The overall average error was 7.31% that predicted
This reduces rubbing of abrasive particles on the workpiece sur- by dimensional model. To compare semiempirical dimensional
face, hence decrease in “rise in temperature” registered on the model with empirical model, a regression model was developed
finishing surface. based on L9 experiments of aluminum 6060 using MINITAB
statistical software, version-18. The adequacy of regression
model was examined by coefficient of determination (R2) and it
Validation indicated the accuracy of fit for regression model. In present case,
R2 and adj. R2 are 91.67% and 83.33%, respectively, depicts good
Five confirmatory tests were performed for the validation of implication of regression model shown in Table 8.
dimensional predictive model on aluminum 6060, opting differ-
ent sets of process parameters that was different from L9OA Rise in Temp ¼  4:00  1:667P1 þ 0:2000P2
experiments for developing the dimensional model. Predictive þ 0:4167P3 þ 0:00333P4
results show that dimensional model had good prediction rate
with 7.31% average error in respect to experimental results.
Advantage of the dimensional model was that it considered
both process parameters and workpiece material properties. The
Table 7. Validation of predictive temperature rise model.
effect of material properties on “rise in temperature” was estab-
“Rise in temperature” (°C)
lished. It was established that semiempirical dimensional tem-
Trial no. P1 P2 P3 P4 Dimensional Experimental % error
perature model was very much capable of predicting an accurate
1 1.0 20 18 300 7.5098 7.0 07.28
“rise in temperature” during surface finishing of MAF process. 2 1.5 20 14 200 3.8205 4.0 04.48
The comparison of confirmatory test results and dimensional 3 2.0 20 18 200 4.8981 5.0 02.03
model predictions of “rise in temperature” are given in Table 7. 4 1.0 25 14 200 5.4622 6.0 08.96
5 2.0 25 18 300 5.1705 6.0 13.82
Table 7 lists the percentage error between the experimental results
1282 R. K. SINGH ET AL.

Table 8. Regression of “rise in temperature” (aluminum 6060).

R-sq R-sq (adj)
91.67% 83.33%
Source DF Seq SS Contribution Adj SS Adj MS F value P value
Regression 4 27.5000 91.67% 27.5000 6.8750 11.00 0.020
Error 4 2.5000 8.33% 2.5000 0.6250
Total 8 30.0000 100.00%
*Seq SS = sequential sum of square, Adj SS = adjusted sum of square, Adj MS = adjusted mean of square

Figure 6 shows the comparison between confirmatory test and the model variables behavior were same as aluminum
results and the predicted results of both “regression model 6060.
and dimensional model.” The graph shows the same variation
in “rise in temperature” with respect to the experimental pffiffiffi !1:6506
1 V α
results and the “regression and dimensional results” for dif- T ¼ ð0:00009997Þ  pffiffiffiffiffiffiffi
ferent trial runs. The results showed good agreement between α kρe
 3=2 0:6844

experiments results and the “regression and dimensional LρCρ wρCρ2
0:0018

results” of rise in temperature. However, regression model  pffiffiffi 

k α k2
predicted more accurate results compared to dimensional  0:0244
Nkα
model with experiments. The graph evidently shows the  (13)
both dimensional model and regression model were capable ρCρ2
in predicting the “rise in temperature” with respect to experi-
The “rise in temperature” in mild steel was more than alumi-
ments for different the trial conditions with good accuracy.
num 6060 because the thermal conductivity (80.5 W/m-K)
Finally, to explore the adaptability of developed
and specific heat capacity (0.45 J/g°K) of mild steel is lower.
Buckingham-π dimensional model to other materials, dimen-
Maximum 14°C raise in temperature was registered for trial
sional model was implemented to predict the “rise in tem-
no. 3 with 18 V voltage, 25 g abrasive weight, 1 mm working
perature” in mild steel microfinished via MAF process. The
gap, and 200 rpm rotational speed, that is, shown in Table 9.
value of dimensionless constant K and values of exponents
The raise in temperature was higher for the higher voltage and
ϕ,φ,γ, and δ were founded by nonlinear approximation of
abrasive weight as it was in aluminum 6060. Lower raise in
a supplementary surface temperature experiment performed
temperature for higher working gap and rotational speed.
on mild steel. Values of K,ϕ,φ,γ, and δ were 0.00009997,
Process parameters behavior with “rise in temperature” were
1.6506, −0.6844, 0.0018, and −0.0244, respectively, were
governed by the exponents of the dimensional products. The
obtained by writing a mathematical code in MATLAB soft-
prediction rate of mild steel dimensional model with experi-
ware – version R2014a using the developed dimensional equa-
mental value was good with % error difference in range of
tions – are present in Eq. (13). The developed model was very
−5.02% to +13.02%.
adaptable to predict the “rise in temperature” in mild steel

8
Rise in temperature (oC)

3 Dimensional
Experimental
2 Regression

0
1 2 3 4 5
Dimensional 7.5098 3.8205 4.8981 5.4622 5.1705
Experimental 7 4 5 6 6
Regression 6.8333 4.1667 4.8333 5.8333 6.1667

Figure 6. Validation of predicted results with experimental results in aluminum 6060.

 MATERIALS AND MANUFACTURING PROCESSES 1283

Table 9. Mild steel predictive temperature rise model. experiments results and the “regression and dimensional”
“Rise in temperature” (°C) results. In this case, dimensional model predicted more accu-
Trial no. P1 P2 P3 P4 Dimensional Experimental % error rate results compared to regression model with experiments.
1 1.0 20 10 100 7.001543 7 0.022% The graph clearly shows the both dimensional model and
2
3
1.0
1.5
25
25
14
18
200
100
12.00113
14.00245
12
14
0.009%
0.017%
regression model were capable in predicting the “rise in tem-
4 2.0 20 18 200 11.30255 10 13.02% perature” with respect to experiments for different the trial
5 2.0 30 14 100 7.597745 8 −5.028% conditions with good accuracy.
The developed dimensional model is compatible with dif-
ferent type of materials. The flexibility of dimension model is
Rise in Temp ¼ 4:08  4:00 P1 þ 0:367 P2 substantiated with result of mild steel finished via MAF pro-
þ 0:708 P3 þ 0:00833 P4 cess. It is observed that the mild steel dimensional model
depicted the same characteristic of material properties and
process parameters to the “rise in temperature” similar to
Again, a regression model was developed using mild steel the aluminum 6060-dimensional model. It is evident from
as workpiece with the help of MINITAB statistical software, Fig. 6 and Fig. 7, that opting different workpiece material
version-18. The adequacy of regression model was examined does not impact the features of dimensional products in
by coefficient of determination (R2) and it indicated the dimensional model and predict the same behavior of “rise in
accuracy of fit for regression model. In present case, R2 and temperature” for different workpiece, in this present case
adj. R2 are 92.79% and 85.58%, respectively, depicts good aluminum 6060 and mild steel.
implication of regression model shown in Table 10. Figure 7 Validation results confirms that proposed methodology is
shows the comparison between experimental results and the capable in predicting “rise in temperature” irrespective to work-
predicted results of both “regression model and dimensional piece material with good prediction rate. The proposed metho-
model” using mild steel. The graph shows the same difference dology can advantageously replace the empirical model such as
in “rise in temperature” with respect to the experimental regression analysis for the prediction of process parameters.
results and the “regression and dimensional results” for dif- Figure 6 and Figure 7 shows that dimensional model prediction
ferent trial runs. The results showed good agreement between results similar to the regression model with respect to the

Table 10. Regression of “rise in temperature” (mild steel).

R-sq R-sq (adj)
92.79% 85.58%
Source DF Seq SS Contribution Adj SS Adj MS F value P value
Regression 4 96.500 92.79% 96.500 24.125 12.87 0.015
Error 4 7.500 7.21% 7.500 1.875
Total 8 104.000 100.00%
*R-sq = coefficient of determination, R-sq(adj) = adjusted determination coefficient

18

16
Rise in temperature (oC)

14

12

10

6 Dimensional

Experimental
4
Regression
2

0
1 2 3 4 5
Dimensional 7.001543 12.00113 14.00245 11.30255 7.597745
Experimental 7 12 14 10 8
Regression 7.1667 12.6667 13.6667 9.6667 8.1667

Figure 7. Validation of predicted results with experimental results in mild steel.

1284 R. K. SINGH ET AL.

experimental results. Dimensional model present better under- magnetic abrasive brush. Electromagnet rotational
standings about the finishing behavior of MAF process incor- speed and working gap are less-important factors. The
porating the contribution of process parameters along with the “rise in temperature” is affected by interaction between
signification of the material properties. Role of material proper- the different process parameters.
ties in MAF finishing process are important as finishing accu-
racy, material removal rate, surface roughness, and surface
temperature depends on the material properties.
ORCID
Conclusion Rajneesh Kumar Singh http://orcid.org/0000-0002-5823-0939

Surface microfinishing of aluminum 6060 using MAF process

is very much suitable, since the experimental investigations
References
reveal that it is a low-temperature finishing process and
makes finishing material thermally stable. Due to low fric- [1] Yan, B.; Chang, G.; Chang, J.; Hsu, R. Improving Electric
tional heat generated on surface of aluminum 6060, make it Discharge Machined Surface Using Magnetic Abrasive Finishing.
free from thermal stress and there are minimal chances of any Mach. Sci. Technol. 2004, 8, 103–118. DOI: 10.1081/MST-
120034246.
microstructural changes in the surface which does not affect [2] Girma, B.; Joshi, S. S.; Raghuram, M. V. G. S.;
the service life of body parts manufacture with it. Further, to Balasubramaniam, R. An Experimental Analysis of Magnetic
predict the role of process parameters along with physical Abrasive Finishing of Plane Surfaces. Mach. Sci. Technol. 2006,
properties of aluminum 6060 on surface temperature, semi- 10, 323–340. DOI: 10.1080/10910340600902140.
empirical dimensional model is established using [3] Jain, V. K.; Jayswal, S. C.; Dixit, P. M. Modelling and Simulation
of Surface Roughness in Magnetic Abrasive Finishing Using
Buckingham’s π theorem. The developed dimensional model Non-Uniform Surface Profile. Mater. Manuf. Processes. 2007, 22,
can be agreeably used to predict the surface temperature rise 256–270. DOI: 10.1080/10426910601134096.
in mild steel. Further, this semiempirical model can be [4] Jain, V. K.; Singh, D. K.; Raghuram, V. Analysis of Performance of
employed for exploring the surface roughness of the finished Pulsating Flexible Magnetic Brush (P-FMAB). Mach. Sci. Technol.
surfaces as well as for surface hardness. Finally, present sur- 2008, 12, 53–76. DOI: 10.1080/10910340701883538.
[5] Wang, A. C.; Tsai, L.; Liu, C. H.; Liang, K. Z.; Lee, S. J. Elucidating
face temperature analysis has led to the following conclusions: the Optimal Parameters in Magnetic Finishing with Gel Abrasive.
Mater. Manuf. Processes. 2010, 26, 786–791. DOI: 10.1080/
● “Rise in temperature” of aluminum 6060 surface in 10426914.2010.505620.
MAF process is low and the highest 9°C “rise in tem- [6] Hung, C. H.; Ku, W. L.; Yang, L. D. Prediction System of
perature” was achieved for trial condition 3. Reason for Magnetic Abrasive Finishing (MAF) on the Internal Surface of
a Cylindrical Tube. Mater. Manuf. Processes. 2010, 25(12),
a low “rise in temperature” is that aluminum 6060 has 1404–1412. DOI: 10.1080/10426914.2010.499578.
a good thermal conductivity which conduct heat quickly [7] Mulik, R. S.; Pandey, P. M. Mechanism of Surface Finishing in
from the microfinished workpiece surface. Ultrasonic-Assisted Magnetic Abrasive Finishing Process. Mater.
● The developed semiempirical dimensional model using Manuf. Processes. 2010, 25, 1418–1427. DOI: 10.1080/
Buckingham-π theorem has successfully predicted the 10426914.2010.499580.
[8] Judal, K. B.; Yadava, V.; Pathak, D. Experimental Investigation of
“rise in temperature.” Dimensional model is derived Vibration Assisted Cylindrical–Magnetic Abrasive Finishing of
from thermal and physical properties of the workpiece Aluminium Workpiece. Mater. Manuf. Processes. 2013, 28(11),
material that had an impact on surface temperature in 1196–1202. DOI: 10.1080/10426914.2013.811725.
MAF process. [9] Judal, K. B.; Yadava, V. Experimental Investigations into
● Predictive Buckingham-π model of the “rise in tempera- Electrochemical Magnetic Abrasive Machining of Cylindrical
Shaped Nonmagnetic Stainless-Steel Workpiece. Mater. Manuf.
ture” is validated by conducing confirmatory tests on Processes. 2013, 28(10), 1095–1101. DOI: 10.1080/
aluminum 6060 and there has been a good agreement 10426914.2013.792427.
with an average error of 7.31%. Dimensional model is [10] Amineh, S. J.; Tehrani, A. F.; Mosaddegh, P.; Mohammadi, A.
compared with regression model with good accordance. A Comprehensive Experimental Study on Finishing Aluminum
● Predictive Buckingham-π model of “rise in temperature” Tube by Proposed UAMAF Process. Mater. Manuf. Processes.
2015, 30(1), 93–98. DOI: 10.1080/10426914.2014.965319.
is easily adaptable in predicting surface temperature rise [11] Sun, X.; Zou, Y. H. Development of Magnetic Abrasive Finishing
in mild steel microfinished via MAF process and the Combined with Electrolytic Process for Finishing SUS304
process variables shows the same behavior as in alumi- Stainless Steel Plane. Int. J. Adv. Manuf. Technol. 2017, 92
num 6060. This predictive model can be easily applied (9–12), 3373–3384. DOI: 10.1007/s00170-017-0408-9.
to other materials such as mild steel. [12] Misra, A.; Pandey, P. M.; Dixit, U. S. Modeling and Simulation of
Surface Roughness in Ultrasonic Assisted Magnetic Abrasive
● Buckingham-π model reveals the significant role of Finishing Process. Int. J. Mech. Sci. 2017, 133, 344–356. DOI:
thermo-physical properties of workpiece material. 10.1016/j.ijmecsci.2017.08.056.
Thermal conductivity (209 W/m-K) and thermal expan- [13] Misra, A.; Pandey, P. M.; Dixit, U. S.; Roy, A.; Silberschmidt, V. V.
sion (23.4 K−1) are most important thermal factors, Multi-Objective Optimization of Ultrasonic-Assisted Magnetic
followed by specific heat capacity (0.91 J/g°K) of alumi- Abrasive Finishing Process. Int. J. Adv. Manuf. Technol. 2018.
DOI: 10.1007/s00170-018-3060-0.
num 6060. [14] Zou, Y. H.; Xie, H. J.; Dong, C. W.; Wu, J. Z. Study on Complex
● The most influential process parameters affecting “rise Micro Surface Finishing of Alumina Ceramic by the Magnetic
in temperature” are voltage andabrasive weight in Abrasive Finishing Process Using Alternating Magnetic Field.
 MATERIALS AND MANUFACTURING PROCESSES 1285

Int. J. Adv. Manuf. Technol. 2018, 97(5–8), 2193–2202. DOI: [23] Kato, T.; Fujii, H. Temperature Measurement of Workpiece in
10.1007/s00170-018-2064-0. Surface Grinding by PVD Film Method. J. Manuf. Sci. Eng. ASME.
[15] Shinmura, T.; Takazawa, K.; Hatano, E.; Matsunaga, M. Study on 1997, 119, 689–694. DOI: 10.1115/1.2836810.
Magnetic Abrasive Finishing. Ann. CIRP. 1990, 39(1), 325–328. [24] Ueda, T.; Sato, M.; Hosokawa, A.; Ozawa, M. Development of
DOI: 10.1016/S0007-8506(07)61064-6. Infrared Radiation Pyrometer with Optical Fibres: Two Colour
[16] Yamaguchi, H.; Shinmura, T. Study of the Surface Modification Pyrometer with Non-Contact Fibre Coupler. Ann. CIRP. 2008, 57,
Resulting from an Internal Magnetic Abrasive Finishing Process. 69–72. DOI: 10.1016/j.cirp.2008.03.056.
Wear. 1999, 225(229), 246–255. DOI: 10.1016/S0043-1648(99) [25] Tsai, K. M.; Wang, P. J. Semi-Empirical Model of Surface Finish
00013-7. on Electrical Discharge Machining. Int. J. Mach. Tools Manuf..
[17] Singh, D. K.; Jain, V. K.; Raghuram, V. Parametric Study of 2001, 41,1455–1477.
Magnetic Abrasive Finishing Process. J. Mater. Process. Technol. [26] Patil, N. G.; Brahmankar, P. K. Determination of Material
2004, 149(1–3), 22–29. DOI: 10.1016/j.jmatprotec.2003.10.030. Removal Rate in Wire Electro-Discharge Machining of Metal
[18] Singh, D. K.; Jain, V. K.; Raghuram, V. On the Performance Matrix Composites Using Dimensional Analysis. Int. J. Adv.
Analysis of Flexible Magnetic Abrasive Brush. Mach. Sci. Manuf. Technol. 2010, 51, 599–610. DOI: 10.1007/s00170-010-
Technol. 2005, 9, 601–619. DOI: 10.1080/10910340500398217. 2633-3.
[19] Ahmad, S.; Gangwar, S.; Yadav, P. C.; Singh, D. K. Optimization of [27] Mulik, R. S.; Srivastava, V.; Pandey, P. M. Experimental
Process Parameters Affecting Surface Roughness in Magnetic Abrasive Investigations and Modelling of Temperature in the Work-Brush
Finishing Process. Mater. Manuf. Processes. 2017, 1532–2475. Interface during Ultrasonic Assisted Magnetic Abrasive Finishing
[20] Hou, Z. B.; Komanduri, R. Magnetic Field Assisted Finishing of Process. Mater. Manuf. Processes. 2012, 27(1), 1–9. DOI: 10.1080/
Ceramics – Part III: Thermal Model. J. Tribol. ASME. 1998, 120, 10426914.2010.515647.
660–667. DOI: 10.1115/1.2833763. [28] Shen, B.; Xiao, G.; Guo, C.; Malkin, S.; Shih, A. J. Thermocouple
[21] Kumar, G.; Yadav, V. Temperature Distribution in the Workpiece Due Fixation Method for Grinding Temperature Measurement. J. Manuf.
to Plane Magnetic Abrasive Finishing Using FEM. Int. J. Adv. Manuf. Sci. Eng. ASME. 2008, 130, 051014. DOI: 10.1115/1.2976142.
Technol. 2008, 41, 1051–1058. DOI: 10.1007/s00170-008-1557-7. [29] Phadke, M. S.;. Quality Engineering Using Robust Design, 11h ed.;
[22] Xu, X.; Malkin, S. Comparison Methods to Measure Grinding Prentice Hall: New Jersey, 1995.
Temperature. J. Manuf. Sci. Eng. ASME. 2001, 123, 191–195. [30] Langhaar, H. L.;. Dimensional Analysis and Theory of Models;
DOI: 10.1115/1.1369358. Wiley: New York, 1957.